Low-level blast exposure induces chronic vascular remodeling, perivascular astrocytic degeneration and vascular-associated neuroinflammation

Abstract

Cerebral vascular injury as a consequence of blast-induced traumatic brain injury is primarily the result of blast wave-induced mechanical disruptions within the neurovascular unit. In rodent models of blast-induced traumatic brain injury, chronic vascular degenerative processes are associated with the development of an age-dependent post-traumatic stress disorder-like phenotype. To investigate the evolution of blast-induced chronic vascular degenerative changes, Long-Evans rats were blast-exposed (3 × 74.5 kPa) and their brains analyzed at different times post-exposure by X-ray microcomputed tomography, immunohistochemistry and electron microscopy. On microcomputed tomography scans, regional cerebral vascular attenuation or occlusion was observed as early as 48 h post-blast, and cerebral vascular disorganization was visible at 6 weeks and more accentuated at 13 months post-blast. Progression of the late-onset pathology was characterized by detachment of the endothelial and smooth muscle cellular elements from the neuropil due to degeneration and loss of arteriolar perivascular astrocytes. Development of this pathology was associated with vascular remodeling and neuroinflammation as increased levels of matrix metalloproteinases (MMP-2 and MMP-9), collagen type IV loss, and microglial activation were observed in the affected vasculature. Blast-induced chronic alterations within the neurovascular unit should affect cerebral blood circulation, glymphatic flow and intramural periarterial drainage, all of which may contribute to development of the blast-induced behavioral phenotype. Our results also identify astrocytic degeneration as a potential target for the development of therapies to treat blast-induced brain injury.

Keywords: Animal model; Astrocyte; Blast; Brain; Chronic; Neurovascular unit; Rat; Tight junctions; Vascular; Vascular pathology.

Fig. 1

Micro-CT scanning of brains of blast-exposed and control rats. Blast-exposed and control rats were transcardially perfused with the Brite Vu contrast agent at 48 h, 6 weeks and 13 months after blast exposure. Brains were scanned at a resolution of 7.5 µm using 0.3° rotational steps of view around 360° and three-dimensional reconstructions were prepared with Bruker’s NRecon software before visualization with Bruker’s CTVox 3D visualization software. Representative maximum intensity projection (MIP) images show dorsal view of volume-rendered brain vasculature from blast-exposed (a–c) and control (d–f) rats. Arrow in b shows an apparently displaced basilar artery in this blast-exposed animal. Arrowhead in b indicates a focally hypoperfused region within the cerebellum. Scale bar, 2 mm

Fig. 2

High resolution images of micro-CT scanning of the rat cerebral vasculature. a lateral view of the left middle cerebral artery (MCA, *) and (b), posterior view of the basilar (*) and the posterior cerebral arteries (arrows) from a control rat, 13 months post-sham exposure. Note visualization of large, medium and small vessels including what are likely precapillary arterioles (arrowhead). Scale bar, 1 mm

Fig. 3

Quantification of X-ray high-resolution micro-CT scanning. Parameters were determined using the software Vesselucida Explorer (v2018.1.1, MBF Bioscience LLC, Williston, VT, USA). Shown is total vascular length a, volume b, surface area c, length density d, the number of isolated vessels e, the longest vessel f, the number of branching nodes g, ending vessels h and the estimated total tissue volume i. Data is shown as violin plots. There were no statistically significant differences between the blast-exposed and sham-exposed (control) cohorts at any of the times studied post-exposure

Fig. 4

Acute, subacute and chronic vascular hypoperfusion and disorganization in the brains of blast-exposed rats. Brain micro-CT coronal optical sections (3.75 mm thickness) of blast-exposed and control rats perfused with the Brite-Vu-contrasting agent and reconstructed with the Vesselucida 360 software. a 48 h post-blast; b 6 weeks post-blast; c 13 months post-blast; d–f 48 h-, 6 weeks- and 13 months sham controls, respectively. Optical sections a and d correspond approximately to coordinates interaural 4.06–7.56 mm; b and e to interaural 6.36–9.86 mm; c and f to interaural 4.06–7.56 mm [97]. Note hypoperfused areas in the brains of blast-exposed rats relative to controls. The arrow in a indicates hypoperfusion of the entire left hemispheric cortical and hippocampal areas. Affected areas in b include the somatosensory, visual, motor and retrosplenial cortical regions whose circulation is derived from branches of the middle cerebral artery (MCA). In c blast-induced cerebral disorganization includes focal tears across the amygdala and adjacent cortical regions (arrows), attenuation of arteries involved in posterior circulation (arrowhead) and hypoperfusion of cortical areas (*). Scale bar, 1.5 mm

Fig. 5

Correlation of behavioral traits with blast-induced cerebrovascular lesions in rats exposed to repetitive low-level blast at 13 months post-exposure. Elevated zero maze (EZM, a–c) and fear conditioning (FC, d–f) testing was performed as part of the studies described in ref [99]. EZM testing was performed at 40 weeks post-blast exposure. Shown for EZM are a open arm entries, b time spent in the open arms and c latency to cross between two open arms. Labeled in color are results from selected rats whose cerebral vasculature is illustrated in Figs. 4 (Fig. 4c, blast-exposed, red square; Fig. 4f, control, aqua circle) and 8 (Fig. 8a, b, blast-exposed, green square). FC testing was performed 42 weeks post-blast. Results are shown for the d training phase, e contextual fear memory testing (assessed 24 h after training) and f cued fear memory testing (assessed another 24 h later). Results of selected blast-exposed rats (Figs. 4c, red and 8a, b, green), are compared to the mean ± standard error measurements of blast and control cohorts determined previously [99]. Statistically significant differences between blast-exposed and control cohorts are indicated (*p < 0.05; **p < 0.01, unpaired t tests) [99]

Fig. 6

Astrocytic degeneration in the rat brain 13 months post-blast exposure. Electron micrographs of frontal cortical arterioles from control a and blast-exposed b–f rats. (*) indicates swollen astrocytic feet at different stages of degeneration. Arrow in a indicates a normal astrocytic endfoot. Arrow head in f shows an elongated mitochondria in an endothelial cell. Scale bar, 1 µm a–c, f; 1.5 µm d; 2 µm e

Fig. 7

Perivascular apoptosis in a vessel with enlarged paravascular spaces. A hypothalamic vessel showing perivascular astrocytes (a, GFAP, red) and microglia (b, Iba1, magenta) analyzed by TUNEL staining (c, green), merged d. DAPI (blue). (*) denotes the enlarged paravascular spaces. TUNEL staining of perivascular astrocytes and of astrocyte-associated microglia is indicated by arrows. Scale bar, 25 µm a–d

Fig. 8

Chronic enlargement of paravascular spaces in the brain at 13 months post-blast exposure. Coronal sections from a blast-exposed animal were stained with hematoxylin and eosin a, c, d. The general pathology showing enlarged paravascular spaces is shown (arrows in a). In this animal, the histopathology is mostly concentrated in one hemisphere and involves the amygdala, insula, piriform cortex, ventral thalamus and hypothalamus. In the contralateral hemisphere, pathological changes are limited to the piriform cortex and hypothalamus. Panel b shows the reconstructed cerebral vasculature from a 3.75 mm micro-CT optical section that includes the region shown in a. The vascular reconstruction b mirrors the enlarged paravascular spaces seen in panel a where the hypothalamus, ventral thalamus and, most strikingly, the amygdala and piriform cortex show reduced perfused vessels (arrows) and enlarged paravascular spaces. The lack of vascular tracing (arrows in a, b) suggests hypoperfusion of the corresponding areas and could be a consequence of perivascular astrocytic degeneration. Panel c shows a detached artery (arrow) within an enlarged paravascular space in the ventromedial hypothalamus. Panel d shows a constricted artery (arrow) in the ventral posteromedial nucleus from which the adventia has been detached. Panel e shows an electron micrograph of a cortical arteriole barely attached to the parenchyma through a few degenerating astrocytic endfeet (arrow). Arrowhead in e shows enlarged paravascular space. Panel f shows immunohistochemical analyses of a hypothalamic arteriole with a disrupted smooth muscle layer (α-SMA⁺, green) and detached adventitia (col IV⁺, red); DAPI, blue. Scale bars, 2 mm a, b; 40 µm c, d; 1.5 µm e; 15 µm f

Fig. 9

Loss of vascular collagen type IV in brain regions with enlarged paravascular spaces in rats at 13 months post-blast exposure. Coronal brain sections of a blast-exposed rat were treated with pepsin to unmask the collagen type IV epitopes and stained with rabbit polyclonal antibodies against collagen type IV (red). a Lack of collagen type IV immunostaining (red) in the piriform cortical region exhibiting large paravascular spaces (arrows). Note the normal collagen IV staining (red) in the adjacent areas (asterisk). The inset in a corresponds to a section from a control animal showing normal collagen type IV staining (arrow) of the vasculature in the piriform cortex and neighboring regions. b Normal vascular collagen type IV staining of the hippocampal vasculature in the same blast-exposed animal (coll IV, red; GFAP, green). c, d Loss of collagen IV immunostaining in small vessels associated with hypothalamic enlarged paravascular spaces. GFAP⁺ perivascular astrocytic endfeet remained attached (GFAP, green, arrows). e Loss of collagen type IV in the arterial adventitia in regions with enlarged paravascular spaces (α-SMA, green). Arrows in e show arterial regions devoid of collagen IV (red) around a vascular stricture (*). Also note the loss of smooth muscle (α-SMA⁺, green) around the stricture. f Loss of adventitial collagen IV (arrows) in hippocampal arterioles (coll IV, red; α-SMA, green). DAPI, blue. Scale bar, 140 µm a; 20 µm for b–f

Fig. 10

Reduced GFAP immunoreactivity in brain regions with enlarged paravascular spaces of blast-exposed rats at 13 months post-blast exposure. Coronal brain sections from blast-exposed a, b and control c, d rats were immunostained with antibodies against GFAP (green) and counterstained with DAPI (blue in merged images, b, d). GFAP immunoreactivity was greatly reduced in the thalamus and hypothalamus of blast-exposed animals harboring enlarged paravascular spaces (arrow in a). e, f Abnormally high perivascular astrocytic density in a few large vessels in the entorhinal cortex of blast-exposed rats. Section in e was single-stained with GFAP (green) and f was double-stained for GFAP (green) and collagen type IV without protease pre-treatment (red,). DAPI (blue). Panel f also shows GFAP⁺ (green) cells under the adventitial collagen type IV (red). Scale bars, 320 µm a, b, d, e; 16 µm c, f

Fig. 11

Microglia/macrophage and astrocytic cells in the vascular lumen. a Confocal optical section (0.56 µm) showing an Iba1⁺ microglia/macrophage (arrow) inside the vascular lumen (Iba 1, red; GFAP, green). Panel c shows an astrocytic cell (arrow) and its process inside the lumen of a small vessel (GFAP, green; coll IV, red). Panels b and d correspond to three-dimensional stack reconstructions of the fields shown in panels a and c. DAPI, blue. Scale bar, 15 µm for a, b; 20 µm for c, d

Fig. 12

Chronic proteolysis of collagen type IV in the brains of blast-exposed rats. Unmasking collagen type IV epitopes in the vasculature of normal mature adult rodent brain for immunohistochemical detection requires protease pretreatment [37]. Shown are micrographs of collagen type IV-immunostained sections of brains of control and blast-exposed rats at 13 months post-blast exposure without protease treatment. a, b, Vasculature in the somatosensory cortex of control a and blast-exposed b rats immunostained for collagen type IV (red). Note the absence of collagen IV staining in the section from the control rat a. c A small vessel immunostained for collagen type IV (red) showing MMP-9 vescicles (green, arrow). d–f Collagen type IV (red) and GFAP (green) immunoreactivity of the somatosensory cortical vasculature of blast-exposed rats in the absence of protease pretreatment. Note the tortuosity of the vessel shown in e and the neuronal expression of GFAP (arrows) in this region f. DAPI, blue. Scale bar, 140 µm for a, b; 20 µm for c–f

Fig. 13

Metalloproteases in the brain of blast-exposed rats. Enriched brain vascular fractions were prepared from control and blast-exposed rats at 6 weeks post-exposure. Protein extracts (20 µg) from the enriched vascular fractions were analyzed by gelatin zymography. Panel a shows vascular upregulation of gelatinase activity associated with MMP-2, MMP-9 and their precursor proteins in the blast-exposed animal (B) as compared to control (C). Panels b, c show a brain arterial cross-section from a blast-exposed rat stained with antibodies against α-SMA (b, green) and MMP-9 (c, red). Panel d shows MMP-9 expression (green) in brain arterial smooth muscle of a blast-exposed rat at 13 months post-exposure. Panel e showns the same vessel in d stained with collagen IV (red). Brain sections of blast-exposed animals were stained with anti-rat IgG (red) and counterstained with DAPI (blue) f, g. Leakage of IgG could be seen in perivascular areas within the corpus callosum f or the entorhinal cortex g. Scale bars, 20 µm b, c 10 µm d, e 100 µm f; 300 µm g

Fig. 14

Blast-induced thickening of the basal lamina surrounding the brain vasculature at 13 months post-blast exposure. Electron micrographs of small vessels in the frontal cortex of blast-exposed a and control animal b. Note that in the vessel from the blast-exposed animal a compared to control b, there is enhanced thickness of the vascular basal lamina (arrows), a reduction of luminal space, swelling of the endothelial nucleus, and reduction of astrocytic endfeet contacts. The endothelial membrane in contact with the lumen is distorted and has an irregular "wavy" pattern. Also, note that in panel a the cytosol of the remaining astrocytes lack the ribosomal density present in the normal astrocytic cytoplasm of control vessels b. Scale bar, 1 µm

Fig. 15

Perivascular microglia and vascular degeneration. a–f Electron micrographs of a perivascular microglial cell (M) adjoining a degenerating arteriole. Perivascular astrocytic feet (As, arrow) exhibit lack of ribosomes and altered mitochondria including swollen mitochondria (sm, panel c), mitochondria with disorganized cristae (arrowhead in c) and completely dystrophic mitochondria (* in c). The intimal endothelial cells (EC, panels a–c) are partially dislodged from the basal lamina with free processes protruding into the lumen. The abnormal endothelial cells exhibit a high density of small osmiophilic vescicles. Note the lack of definition of the basal lamina between endothelial and smooth muscle (SM) cells (arrow in c). The perivascular microglial cell a, b, d–f shows lysosomal cholesterol crystals (cc in panels b and d), a lipofuscin body (lf in panels b and c) and enlarged endoplasmic reticulum cisternae (er) characteristic of the “dark” microglial phenotype e. Arrowheads in panel d show abnormal enhanced interstitial spaces in the neuropil neighboring the perivascular microglial cell. Arrowhead in panel f indicates remnants of an astrocytic foot process next to a microglial process. m, mitochondria; tj, tight junction. Scale bars, 2.5 µm a; 1 µm b; 0.2 µm c; 0.4 µm d, e; 0.8 µm f

Fig. 16

Perivascular microglia in the brain of blast-exposed animals. Panels a–c show clustered Iba 1⁺ microglia (red, arrow in b) with amoeboid morphology in the perivascular area of a tortuous large vessel (pericallosal artery) in the corpus callosum of a rat brain at 13 months post-blast exposure. Panel d shows an orthogonal projection of a section through a thalamic artery stained for perivascular astrocytes (GFAP⁺, green) and microglia (Iba1⁺, red). Co-localization of astrocytic GFAP⁺ material within a perivascular microglial cell, suggests phagocytic ingestion of perivascular astrocytes by microglia. Scale bar, 100 µm a; 40 µm b; 20 µm c, d

Fig. 17

Perivascular inflammation in the brain of a blast-exposed animal. Panel a shows a brain region with patches of clustered activated microglia (Iba1⁺, red) next to a vascularized region in between the deep cerebral white matter, choroid plexus and hippocampal fimbria. Insert in a shows higher magnification of a 0.56 µm optical section of the area indicated by arrow. Note the tortuous vessel (arrow). Panel b denotes higher magnification of an artery (identified by * in panel a) with perivascular activated microglia (Iba1⁺, red). Panels c, d show GFAP⁺ intracellular material (green) in perivascular activated microglia (Iba1, red; GFAP, green). Scale bar, 200 µm a; 10 µm b–d

Fig. 18

Perivascular microglial activation. MHCII expression is associated with microglial activation. Shown are perivascular amoeboid Iba1⁺ microglia (red). Some microglia are also expressing MHCII (green) and extend transcellular processes across the nuclei of Lochkern-like cells. a, Merged images, b Iba1 staining (red); c MHCII staining (green); and d DAPI staining (blue). Insert in d shows an enhanced view of a Lochkern-like cell. Arrows indicate the location of the Lochkern-like cells with hollow nuclei and the transcellular microglial processes (Iba1⁺ and MHCII⁺). Cellular blebs indicate that MHCII⁺ microglia appear to be undergoing apoptosis. Scale bar, 15 µm a–d; 5 μm insert in d

Fig. 19

Summary timeline of cerebral pathological alterations in a rat model of blast-induced mild TBI. Shown are observations in blast-exposed animals from acute, subacute and chronic cohorts, all of which received 3 × 74.5 kPa exposures [, –41, 63, 66, 67]. Acute, 24-72 h post-blast; Subacute, 6 weeks post-blast (normal behavioral phenotype); Chronic, 16–52 weeks post-blast, (PTSD-like behavioral phenotype)